IEEE LATIN AMERICA TRANSACTIONS, VOL. 13, NO. 5, MAY 2015
1429
Wind Resource Potential in Los Taques Venezuela F. M. Gonzalez-Longatt, Senior Member, IEEE Abstract— The Venezuelan government has established more aggressive policies and incentives for renewable energy resources in recent time, especially in terms of wind power. Although several academic efforts to make publically available wind energy resource data in Venezuela, there is a lack of information in terms of local wind resource putting in risk development in areas where potential is good enough for commercial exploitation. The objective of this paper is to presents a very comprehensive wind resource assessment at Los Taques, Venezuela based on on-site observation anemometry. This is unique paper because it is the first ever wind energy assessment in Los Taques using hourly data recorded during three years in an on-site ground weather station contrary to studies based on daily values based on radar or satellite data. The applied methodology has been developed based on the characteristic of the data obtained from the on-site anemometry. Results of wind energy assessment and evaluations on a 100 MW wind farm shows the wind energy resource available in Los Taques is enough for commercial use and the results. Keywords— Wind data, Wind Energy potential, Wind power generation, Venezuela.
T
area of Santa Cruz de los Taques (or Los Taques as known).This paper is a first effort make publically available information about the wind resource potential available at Los Taques-Venezuela, it will allow to local population a valuable insight into the wind resource, its potential development, and its value to a utility utilization of individual use. The objective of this paper is to presents a very comprehensive wind resource assessment at Los Taques, Venezuela based on on-site observation anemometry. Section II describes briefly the analysis method for the assessment in the study area whilst Section III to V present the results gathered and a discussion of their significance. Data used in this paper is based on the available wind data measurement from on-site observation anemometry. From the results of this paper, Los Taques is identified as suitable site for the wind energy exploitation in Venezuela. Conclusions of this paper suggest further site-specific investigations should be conducted evaluating economical of potential wind energy development.
I. INTRODUCTION
HE BOLIVARIAN Republic of Venezuela is a country which has the largest electricity consumption in South America (4,018 KWh/year per capita) and electrical power system provide electricity to 95% Venezuelan population [1]. The demand peak value varies between 16,500 MW and 18,200 MW depending on seasonal conditions [2], [3]. Electricity consumption rises between 4% and 7% per year, and it is expected to increase with the same or higher rate in the next 10 years [2]. Total generation installed capacity is 26,550 MW and the generation mix is 65% hydropower, 32% thermal power plants and 3% distributed energy resources [1]. Although the proven oil reserves in Venezuela are claimed to be one of the largest in the world, more aggressive policies on the use of environmentally friendly electricity generation have begun in recent years in Venezuela. Several academic projects have been reported to promote renewable energy sources installations in numerous areas of Venezuela [4], [5], [6] especially wind power. Several smallscale and off-grid wind power projects have been developed and two utility-scale wind have been installed in mainland Venezuela: La Guajira (25 MW) [7], and La Peninsula de Paraguaná (100 MW) [8]. A wind atlas of Venezuela has been recently published by the author in [1] where several areas have been identified suitable for wind energy projects, including the Paraguaná area where the Paraguaná Wind Farm is installed [8]. However, there is not information, publically available to allow enforce more development of wind energy use in the F. M. Gonzalez-Longatt, Loughborough University, United Kingdom,
[email protected]
Loughborough,
II. METHODOLOGY WIND POTENTIAL ASSESSMENT Wind energy site assessment evaluates the potential for a given site to produce energy from wind turbines. There are several approaches to investigate the wind resource within a given area of land [9] and the preferred approach is defined by objectives of the wind energy program. [10]. However, there is a general consensus as to how wind energy site assessment is performed. The Wind Resource Assessment Handbook [11], Wind Energy – The Facts (Volume 1, Chapter 2) [12], consulting firms [13], and state guidebooks on site assessment [14], all endorse a similar site assessment methodology. As summarized in [15], there are a number of methods for estimating the wind resource of an area [16], [17]. A detailed review of all of these methods is beyond the scope of this paper. Aspects of wind resource evaluation based on measurement only are presented in this paper. This approach has been applied successfully in several locations around the world [18-23]. Figura 1 shows a complete flow chart of the methodology for the wind resource assessment followed in this paper. This methodology has been developed by the author based on the characteristic of the data obtained from the on-site anemometry. This simplify procedure follows a sequence of three steps: (1) data validation, (2) data recovery, and (3) data processing. The main input data for the site wind resource assessment procedure is on-site measurement data, time series, relating to different meteorological parameters: wind speed, wind direction, air temperature, and atmospheric air pressure. The onsite measured data must be validated and processed in order to generate adequate information to allow wind
1430
IEEE LATIN AMERICA TRANSACTIONS, VOL. 13, NO. 5, MAY 2015
resource assessment. The data validation process consists of the inspection of all the collected data for both completeness and integrity as well as the elimination of any erroneous values. There is several validation routines designed to screen each measured parameter for suspect values before they are incorporated into the archived database and used for site analysis. Manually and automatically routines are used for validation purposes in this paper. Details of validation tests are presented in next section. When the data validation step is complete, the data set must be subjected to various data processing procedures to assess the wind resource [24], [25]. This typically involves performing calculations on the data set, as well as binning (sorting) the data values into useful subsets based on your choice of averaging interval. The processed data can be analysed in many different ways. However, there is a general consensus about the use of descriptive statistic in preliminary assessment and resource description to quantitatively describing the main features of wind resource. In the following subsections, the treatments used for the processing of valid data used in this paper are presented. Measurements: Wind Speed, Wind Direction Input Data Screening Data Verification
Step 1. Data Validation
Calculation Data Recovery Rate (DDR) DDR < ε ?
Step 2. Data Recovery Data Data unusable
Estimation of Missing data
Valid Data Wind Speed Wind Direction Wind Power Density
Step 3. Data Processing
Output Wind Speed Characteristics
Figura 1. Flow Chart of Methodology for Processing of Wind Data Set and Wind Resource Assessment.
A. Descriptive Statistic Wind Energy The procedure of determining if a site is suitable for wind power production requires convincing statistical information describing the long-term behaviour of the wind resource. Several statistical indicator of wind energy resource are used in the specialized literature. Average speed indicates the overall wind potential at a given site, expected wind speed for a given time interval (first central moment). The variability of wind speed in a given time-series is calculated by the standard deviation (σm). It indicates the mean amplitude of temporal (or spatial) wind fluctuations (square root of the variance). Probability density functions (PDF) such as the Weibull or Rayleigh functions are usually used to determine the wind speed distribution of a windy site for a period of time. Wind speed distributions are used as stochastic representation of the wind resource at the studied site, Weibull probability density function is used in this paper and the maximum likelihood
method is used to obtain distribution parameters [9],[26]. B. Energy Output and Wind Power Density (WPD) The process to estimate the energy output of a wind turbine in the measured wind regime consists of four main steps. First, it estimates the wind speed at the hub height of the wind turbine for each time record in the data set (time step). Second, it uses the hub height wind speed and air density for each time step to estimate the gross power output of the wind turbine for each time [9]. Third, it finds the overall mean and the mean for each month of the gross power output, and multiplies this value by the overall loss factor to calculate the mean net power output, for each month and for the entire data set. Finally, it multiplies the mean net power output by the number of hours in a year (8760) to find the net annual mean energy production. Similarly, it multiplies the monthly mean net power outputs by the number of hours in each month to find the net monthly mean energy production. Full details of this methodology are found several publications [9], [16]. Apart from wind speed, the kinetic energy content of the atmosphere also depends linearly on air density [27]. Nearsurface air density is defined as the mass of a quantity of air divided by its volume. It can be calculated using the ideal gas law. III. DATA SOURCE The data used in this paper was obtained from the meteorological station of the Josefa Camejo Airport (IATA: LSP), located at coordinates of 11°46′07″N and 70°08′09″W at 23m above sea level. This site is found to be the most suitable information source in the area for developing the preliminary wind energy assessment of Los Taques as there are no obstacles around the measurement area so it is directly open to the Venezuelan Gulf to the west. The collected data covers three years period, from 1st January 2008 to 31st December 2010 (1096 days). This station recorded the wind speed, wind direction, temperature, humidity and atmospheric pressure on an hourly basis. The terrain in the surrounding area is relatively flat and suitable for wind power development with very low surface roughness conditions. A three cup anemometer and a wind vane are mounted individually on cross arm supported by single tubular pole, which was erected in July 2007. TABLE I. NOMINAL CHARACTERISTICS AND SPECIFICATIONS OF THE MEASURING EQUIPMENT AT ON-SITE WEATHER STATION. Measurement Accuracy Resolution range 0-160 mph ±0.15 mph 0.05 mph Anemometer 0-71 m/s or 1% Wind vane 0-360° ±2% < 1.0° Thermometer (-40)-(+60) °C ±0.1ºC 0.05 ºC Hygrometer 0-100% RH ±1.5% RH 0.05% Barometer 500-1100hPa ± 0.05hPa 0.01 hPa
Temperature, relative humidity and atmospheric pressure data are obtained from a thermometer, a hygrometer and a barometer, respectively. A data logger is connected to the sensors on the mast to collect data in time series. Table I shows the technical specification of the main measurement
GONZALEZ-LONGATT : WIND RESOURCE POTENTIAL IN
devices installed at the weather station and all wind sensors are mounted according to the World Meteorological Organization (WMO) standard [28]. IV. DATA VALIDATION AND RECOVERY Three-year data set on hourly basis is used in this paper. This extensive time series has been validated manually and automatically (using computer-based techniques). Initially is validated automatically by taking advantage of the power and speed of computers and manually validate where more analysis is required. The validation process includes validation test of wind speed and direction data series in order to verify a normal operation band (wind speed between 0.0 and 25.0 m/s, and wind direction 0°-360°). The data screening is used for the data series validation, filter by flag is used to remove questionable or erroneous, e.g. data like prolong calm timeperiods. Results of validation process showed that the data series of wind speed and direction are inside the normal operation band. In addition, it is not necessary to apply any shifting to the time series of data based on the criteria concerning the maximum expected change of variable over time. Missing data is a common problem in statistical analysis. Rates of less than 1% missing data are generally considered trivial, from 1 to 5 % are manageable. However, from 5 to 15% requires sophistically methods to handle and more than 15% may severely impact any kind of interpretation [29]. Missing data is a source of uncertainty in wind energy resource assessment studies. Several publication recommend that missing data should not exceed 10% [10], and this paper assumed 10% value as maximum. The completeness of the collected data is assessed using the Data Recovery Rate (DDR), it is a measure of the amount of wind data successfully captured by the data logger and is expressed as a percentage of the data records available in a given period of time [10], [30]: Data Recovery Rate (DDR) =
1431
Results of data imputation and the main statistical index are shown in Table II. The statistical measure, Root Mean Square Error (RMSE) is used to indicate how closely the predicted values match the measured values. Results show RMSE below 0.5% for all times-series considered. TABLE II. RESULTS OF WIND SPEED AND DIRECTION DATA IMPUTATION Wind Year 2008 2009 2010 Iteration Number 3 8 6 Speed RMSE (%) 0.365 0.325 0.421 MSE in Estimated Data ( m/s) 2.085 1.139 1.728 Iteration Number 14 6 12 RMSE (%) 0.395 0.453 0.482 Direction 7.624 7.415 8.676 MSE in Estimated Data (°) RMSE: Root Mean Square Error, MSE: Maximum Standard Error after using the EM algorithm.
V. DATA PROCESSING A. Temperature, Pressure and Air Density Figura 2and Figura 3 show monthly and averaged values of temperature and atmospheric pressure for the site, as assessed during the observation time. The temperature average registered is 27.55°C with the minimum diary is 14°C, which is registered in May and maximum diary of 39°C during August. The maximum and minimum monthly mean temperatures are 32.6°C in September and 22°C in August.
Data records collected × 100% Data records available
(1) where records collected is the difference between the data records possible and number of invalid records. The on-site measuring period shall be at least one year and the data recovery rate more than 90 % in order to ensure the quality of the wind energy resource assessment [31]. The total data records possible during 3 successfully measured years is estimated at 26304 which results in a total recovery data of 97.23%, and calculated yearly DDR of wind speed is 99.4%, 95.6% and 96.7% on 2008, 2009 and 2010 respectively. Results of data recovery rate of wind direction during are lower than wind speed during the recording period: 98.7%, 95.0% and 96.2%. In this paper, a variation of the expectation maximization (EM) named regularized EM (RegEM) algorithm is used to replace any missing data and therefore complete the data set. MATLABTM implementation of regularization methods is adapted to fit the framework of the EM algorithm, this is the EM Regularization Tools (RegEM) [32].
Figura 2. Monthly and daily statistics of temperature.
The average atmospheric pressure varies between 1008.42 and 1013.51 mbar with yearly mean value of 1011.18 mbar. The maximum daily value is 1019 mbar which is registered in January and minimum daily is 1001.00 mbar during June and July. Mean diurnal profile shows small changes in atmospheric pressure and largest values are expected between
1432
11:00 and 12:00 hours (1012.3 mbar). The site-specific air density is calculated based on on-site measurement of air temperature and atmospheric pressure, the mean values during the observation period is 1.175 kg/m3. The maximum and minimum monthly average air densities are 1.230 kg/m3 in May and 1.124 kg/m3 in August, respectively. The normal environmental conditions are defined by IEC 61400 and it considers air density of 1.225 kg/m3 at sea-level at 15°C to be normal [33], [34]. Average air density values at the observed point are below this standard value during the year, this means that the air density of the site would negatively affect the performance of a wind turbine most of the time. The energy in the wind will be reduced proportionally to the density of air and larger wind turbines are required for the same rated power compared with the conditions specified in the standard.
IEEE LATIN AMERICA TRANSACTIONS, VOL. 13, NO. 5, MAY 2015
average daily high and average daily low, maximum and minimum values. The monthly mean wind speed varies between 4.92 and 11.78 m/s. The maximum value of the mean wind speed occurs in June whilst the minimum value occurred in November and the average speed for the yearly mean is 8.29 m/s.
Figura 4. Monthly statistics and daily profile for wind speed at 50 m height above ground.
Figura 3. Monthly and daily statistics of atmospheric pressure.
B. Wind Speed The change in wind speed with height above ground, wind shear, can be approximated using Prandtl logarithmic law (logarithmic law or log law). This law assumes that the wind speed varies logarithmically with the height above ground [35] and uses the surface roughness (sometimes called surface roughness length or just roughness length) to characterize the wind shear. In this paper, the logarithmic law is used to approximate the wind shear of wind speed data set to height of 50 m and roughness length of 0.0024 m or Roughness Class (RC) of 0.5 is assumed. Those values are representatives of open terrain with a smooth surface, such as concrete runways in airports, mowed grass. A preliminary description of the wind speed at 50 m of the site for the observation period is created using boxplot, as shown on Figura 4 where of five statistical measures: mean,
The wind speed operating range of most horizontal axis wind turbines is defined between cut-in and cut-out wind speeds of about 4 m/s and 25 m/s respectively. It can be noted that the monthly average speeds is over the cut-in during whole year. The average daily high wind speed is 16.25 m/s and occurs in June. The maximum hourly wind speed registration is 22.49 m/s in June and this value is below the cut-out speed of most wind turbines. The wind speeds can be classed as calm (800W/m2) about 21.56% of the total time, 5.17 hours per day, and the WPD is considered poor (
